Analyzing query performance for design optimization

Applies to: Kyvos Enterprise Kyvos Cloud (SaaS on AWS) Kyvos AWS Marketplace

Kyvos Azure Marketplace Kyvos GCP Marketplace Kyvos Single Node Installation (Kyvos SNI)

Kyvos collects statistical information about usage, user interaction patterns, and query response patterns.

You can analyze semantic models and query performance using the Query Analyzer feature.

This functionality provides information about performance statistics, user data, performance trends, and mostly used dimensions and measures. You can also drill down for more details on a specific query. All of this data helps in recommending design enhancements for optimizing the semantic model and query performance. Details such as total query count, average response time, minimum response time, maximum response time, total users, tracking queries by users, most used dimensions, least used dimensions, not used fields, and Top 100 slow running queries, dimensions, and measures used in filtering are also displayed.

You can also analyze live queries from the Activity Monitor.

It is recommended to use these features to analyze slow-performing queries and semantic models.

Example

Consider a case where users report slow queries for a particular semantic model. In such a scenario, you can view the Cumulative Query Analysis dashboard to see an overall picture of the semantic model performance. Say, you found that date filter is getting applied in most of the queries but you have not partitioned the semantic model based on that date. You can then change the design and partition the semantic model on the date, which will help in optimizing query performance.

Consider another case where you find that the semantic model is performing fine, but one particular user query is taking time. In such a case, you can analyze the specific query from the Activity Monitor to see what went wrong.

Related topics